Bringing AI down to earth
There’s a huge buzz around the potential for AI in health and homecare right now. But an immediate focus on the basic digital building blocks is essential if we’re going to benefit from the possibilities of the future...
A study led by Lord Darzi, a former health minister and leading surgeon, imagines ‘bedside robots’ appearing as part of the ‘full automation’ of health and care services. Professionals would be freed up to focus on the value-adding tasks that only humans can do, whilst the country would save billions every year.
At Konnektis, we share this excitement and are strong believers in the potential of technology to improve the quality and efficiency of health and homecare. But we need to overcome the challenges we face today if we are to unlock this exciting future. So what are they?
Let’s look at our own area of home care. Robotics, predictive diagnostics, interpretation of sensor data and many other elements of the home care future that people get excited about all depend on AI. Artificial Intelligence, as the term is most often used, is a product of Machine Learning (or ML). It’s what a computer intelligently does in a particular situation having understood - through a huge amount of data testing and analysis - what is appropriate and will work best.
So, for instance, in homecare an AI programme could alert carers and family members if someone receiving care was on the road to hospitalisation. The programme would have come to this conclusion having analysed thousands of care records and associated hospital admissions and identified the contributory factors that the data revealed as the most influential. The software programme would have noted in this particular situation the warning signs and flagged the likelihood of hospital admission.
In this way a computer may well be able to perceive the significance of a situation where humans might not. An experienced carer may have a hunch but the computer will have evidence for that hunch and be able to apply the insight systematically. It’s now also clear that ML programmes can make connections across mountains of data - between cause and effect, for instance - that a human would never be able to spot.
Sounds good? Yes! But let’s think through what we need in place to get from here to there.
Firstly, we need thousands - ideally, millions - of digitised individual care records and outcomes to support machine learning. These will allow the programme to identify which inputs led to particular outputs in a statistically robust fashion.
Secondly, we need to track individual care processes through digital means on an ongoing basis so interventions can be made using our AI programme. And we need to do this whilst meeting all data protection and privacy requirements.
It’s clear what is the key requirement. The application of both machine learning and the consequent artificial intelligence are entirely dependent on the availability of digital records. Without this we can’t even begin to think about how future technologies can support home care through AI.
The fact is that care in the home is still supported to a significant degree by pen and paper [link to last blog post]. We need to move this critical information onto a digital platform as soon as possible if we’re going to turn our speculations about the future into predictions.
At Konnektis we are fully focused on this crucial process of moving care on to a digital platform. It’s why we created our platform and it’s how our operational teams spend most of their time.
However, what we find interesting - and what the sector will find challenging - is that the key to this innovation is not to be found solely in the technology. Not by any means. Progress actually depends on understanding how people work in care and how their current processes work.
We must never forget that home care is about people, whether the carers or the individuals receiving care. Only having understood care from the perspective of people on the ground can we configure the right digital tools. This means our job is as much about understanding people as it is about tech.
